[
https://issues.apache.org/jira/browse/HIVE-19500?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16487650#comment-16487650
]
Jesus Camacho Rodriguez commented on HIVE-19500:
------------------------------------------------
It is not pleasant to do this. I agree and I trust you... and however, we
decided as a community that no patch will go in without a clean QA run anymore.
That is the only way to keep the build green. If the queue is long, that is
another problem that we need to solve, and making the devs unable to bypass the
ptest queue is the only way they care about making our test runs more stable
and faster. FWIW, there are already patches almost ready to go in and make our
runs faster, e.g., HIVE-19633.
> Prevent multiple selectivity estimations for the same variable in conjuctions
> -----------------------------------------------------------------------------
>
> Key: HIVE-19500
> URL: https://issues.apache.org/jira/browse/HIVE-19500
> Project: Hive
> Issue Type: Sub-task
> Affects Versions: 3.0.0, 3.1.0
> Reporter: Zoltan Haindrich
> Assignee: Zoltan Haindrich
> Priority: Major
> Fix For: 4.0.0
>
> Attachments: HIVE-19500.01.patch, HIVE-19500.02.patch,
> HIVE-19500.03.patch
>
>
> see HIVE-19097 for problem description
> for filters like: {{(d_year in (2001,2002) and d_year = 2001)}} the current
> estimation is around {{(1/NDV)**2}} (iff column stats are available) ....
> actually the source of the problem was a small typo in HIVE-17465
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
